Long-term soil moisture dynamics in a mining region: temporal trends and spatial patterns from GLDAS Noah data

Abstract

This study examines long-term spatiotemporal trends of soil moisture in a mining region using monthly surface soil moisture (0-10 cm) data from the GLDAS Noah Land Surface Model for the period 2000-2024. Temporal trends were analyzed using the non-parametric Mann-Kendall test and Sen’s slope estimator, while seasonal spatial patterns and soil moisture anomalies were used to assess long-term changes. In addition, the Hurst index was applied to evaluate the long-term memory of soil moisture variability. The results indicate a consistent declining trend in soil moisture across all months during the study period. The strongest decreases occur in late autumn and early winter, with November and December showing statistically significant downward trends. Spatial analysis reveals widespread soil moisture reductions across all seasons when comparing conditions in 2000 and 2024. Hurst index results show persistent behavior mainly in autumn and early winter months, indicating strong long-term memory and a tendency for soil moisture deficits during these periods to persist over time. In contrast, spring and summer months exhibit mainly anti-persistent or near-random behavior, reflecting greater short-term variability. Overall, the findings suggest that soil moisture drying in the study area is cumulative and longlasting. This study demonstrates the usefulness of integrating land surface model data with trend and persistence analyses to better understand soil moisture dynamics in mining-affected regions and to support environmental management and land rehabilitation under changing mining conditions.